MISY 5370

Assignment 3

Total Points: 8

Answer the following questions based on the web file named FBS.

Please submit both the 1) word document containing your answers, 2) Excel worksheets containing the relevant tables, and Dendrogram generated after executing hierarchical clustering.

Refer to the clustering problem involving the file FBS described in Assignment 2. Apply hierarchical clustering with 10 clusters using football stadium capacity, latitude, longitude, endowment, and enrollment as variables. Be sure to Normalize input data in Step 2 of the Hierarchical Clustering procedure. Use Ward's method as the clustering method. At step 3, set both Maximum Number of Leaves and Number of Clusters as 10. Make sure "Draw Dendrogram" option is checked. Please create the dendrogram on the HC_Dendrogram worksheet. Copy the dendrogram from Excel to Word (2 points). And

a)draw a horizontal line at the distance 22 and indicate the composition of the clusters segment (3 points).

b)draw a horizontal line at the distance 14 and indicate the composition of the clusters segment (3 points).

Tips: Please read the textbook (complete version) on page 260 to 262, you need to draw a horizontal line and indicate the composition of the clusters segment, like the example provided on page 261 of your textbook.

* If you cannot see the Dendrogram, please read the following solutions: It seems that different people have different layouts under the Data Mining tab. On my computer, I clicked the Data Mining tab, the Dendrogram chart popped up. The version I use does not have the Chart button. But if you have, you may click the Charts button in the Analysis group under Data Mining Tab, the Dendrogram chart should pop-up.

Steps to draw a horizontal line on dendrogram: after the dendrogram was generated in Excel, right-click the dendrogram and select Copy. And then paste it into your word document. Insert a line from the menu "Shapes" from the "Insert" tab.

The purpose of practicingthis assignment is to enhance students' understanding about hierarchical clustering, as well as to constructhierarchical clustering. You don't have to describethe group characteristics of observations. To indicate the composition of the clusters segment, the format like below is fine:

Cluster 1:{the number ofsub-clusters,the number ofsub-clusters,the number ofsub-clusters,...}

Cluster 2: {the number of sub-clusters,the number ofsub-clusters,the number ofsub-clusters,...}

...

Cluster N: {the number ofsub-clusters,the number ofsub-clusters,the number ofsub-clusters,...}